Machine Learning: The next great leap
summary
Argues that machine intelligence is a new way to write software and cites concrete results: Gmail spam rates drop to near zero, Smart Reply ships in Inbox by Gmail and speeds mobile email triage, Google Photos and Translate use on-device neural models and Translate handles 27 languages for sign translation, training moves from 5–6 to 25–30 layer deep neural nets, ecosystem bets include $125M from Google Ventures and 500 startups using $100k cloud credits, cloud tools like DataProc (<90s spinup) and preemptible VMs (≈60% cheaper) are live.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 2:00Google Photos
- 3:00Smart Reply
- 4:00Ecosystem
- 5:00Community
- 7:00Free Apps
- 8:00Google Cloud
- 9:00Outro
related talks
27mKeynote - Google Cloud, Machine Learning
07mMachine learning at Google
59mMachine Intelligence at Google Scale
24mMachine Intelligence gets real
47mBuilding smarter apps with Machine Learning, from magic to reality
23mWhat Will You Do Next?
54mBuilding smarter apps with Machine Learning, from magic to reality
44mBuilding smarter solutions with no expertise in machine learning